On the acquisition of Spanish psych predicates

Abstract

Spanish psych verbs like gustar (‘like’/‘please’) have a non-agreeing dative experiencer that asymmetrically c-commands the agreeing nominative theme (e.g., Cuervo 2003). Intervention accounts (Friedmann et al. 2009) thus predict children will experience difficulties with constructions that involve movement of the nominative-bearing argument past the dative DP. In this study we evaluate this prediction with a corpus study and an experimental study. Results from the corpus study show that children under the age of 7 underuse the DPnom-V-DPdat order (Theme-Verb-Experiencer) with gustar compared to adults, in line with our predictions. In a picture-matching task we tested 4-6-year-olds on d-linked wh-questions with actional and psych verbs. Results reveal that while children display the expected Nom > Dat wh-phrase extraction asymmetry with actional verbs, they show a Dat > Nom asymmetry with psych verbs. Moreover, children perform worse when the two arguments match in number features (i.e., sg-sg) compared to when they mismatch (i.e., sg-pl), but crucially, only in wh-questions that involve intervention. Our results cannot be fully explained under input-based accounts and are most in line with a structural account such as featural Relativized Minimality (Rizzi 2004).